structural integrity of steel hydrocarbon...

1 Copyright © 2014 by ASME

Proceedings of the 10th International Pipeline Conference IPC2014

September 29 – October 03, 2014, Calgary, Alberta, Canada

IPC2014-33210

STRUCTURAL INTEGRITY OF STEEL HYDROCARBON PIPELINES WITH LOCAL WALL DISTORTIONS

Aglaia E. Pournara Dept. of Mechanical Engineering

University of Thessaly, Volos, Greece email: [email protected]

Spyros A. Karamanos Dept. of Mechanical Engineering

University of Thessaly, Volos, Greece

Theocharis Papatheocharis Dept. of Civil Engineering


Philip C. Perdikaris Dept. of Civil Engineering


ABSTRACT Local distortions on pipeline wall in the form of dents or

buckles may constitutea threat for the structural integrity of the

steel pipeline. In the present paper, experimental research

supported by numerical simulationis reported to investigate the

structural integrity of smoothly dented steel pipes. A series of

six (6) full-scale experiments on 6-inch X52 pipes has been

carried out, and numerical simulations have also been

conducted.The dented steel pipesare subjected to cyclic loading

(bending or pressure) in order to estimate their residual strength

and remaining fatigue life. The finite element analysis simulate

the experimental procedure for each type of deformation and

loading case, in order to estimate the local stress and strain

distributions at the dented region. Based on the numerical

results, fatigue life is predicted and compared with the

experimental results. The results of the present study are aimed

at evaluating existing guidelines and methodologies towards

appropriate assessment of local wall distortions in steel

pipelines.

INTRODUCTION

Steel hydrocarbon pipelines may exhibit structural failure

because of damage due to external interference, mechanical

damage, or corrosion [1][2][3]. In the event of pipe wall dent,

caused mainly by external interference, the pipeline may appear

to fulfill its transportation function, provided that the steel

material is adequately ductile and no cracks occur. However,

the damaged area is associated with significant strain

concentrations and, in the case of repeated loading cracks may

develop, leading to fatigue failure. Therefore, the dented area

and the associated strain concentrations may constitute a

possible threat for the structural integrity of steel pipelines [3].

Evaluation of the remaining life of oil and gas transmission

steel pipelines with local wall distortionshas been considered a

crucial issue for maintaining a reliable level of pipeline

operation condition, and has triggered significant amount of

research. Using four different two-dimensional shape functions

to describe dent geometry, Ong [4] calculated local stresses at

the pipe wall of a pressurized elastic pipe with a long smooth

dent and found that the maximum local elastic stress varies

between 14 and 21 times the value of the nominal hoop stress

due to pressure. Ong et al. [5], using finite element simulation,

calculated local elastic stresses at short and smooth dents, in

pressurized dented cylinders. The peak elastic stress in short

dents was found to be significantly less than in long dents. It

was also concluded that the presence of local dents does not

affect the burst capacity of the dented pipeline.

An experimental and numerical investigation of the fatigue

resistance of offshore pipelines with plain (smooth) dents under

cyclic pressure has been reported in [6]. A longitudinal wedge

shaped denting tool was employed to dent 12 in. steel pipe

specimens at depths ranging between 5% and 20% of their

diameter. It was found that plain (smooth) dents with a depth

larger than 5% may result in a reduction of the pipeline fatigue

strength. Buitrago and Hsu [7], using a finite element

simulation, investigated the fatigue response of non-

pressurized tubular members containing relatively smooth dents

and provided stress concentration factors in the form of

parametric equations in terms of geometric parameters for axial

and bending loads.

Recently, the effect of various defects (gouges,

manufacturing or weld defects, corrosion) on the structural

integrity of pipelines has been examined in a joint industry

project [3][8][D], based on existing experimental and numerical

mailto:[email protected]


results. Considering a large number of publications regarding

the ultimate capacity or the fatigue strength of defected or

damaged pipelines, listed in [3][8], that study was aimed at the

enhancement of the current methodologies and the

identification of “gaps” in existing knowledge towards a

“fitness-for-purpose” pipeline assessment. It was found that

very limited data exist for the influence of buckles on the

fatigue resistance of pipelines, while the fatigue capacity due to

bending loads is not yet included.

Das et al. [9] conducted full-scale laboratory tests to

investigate the post-wrinkling ultimate behavior of steel

pipelines. The pipe specimens exhibited extreme ductile

behavior and did not fail in fracture under monotonically

increasing axisymmetric compressive axial loads and

displacements. Fractures developed at the wrinkled region,

however, when a wrinkled pipe specimen was subjected to

cyclic strain reversals due to unloading and loading of primary

loads. More recently, Dama et al. [10]presented experimental

and numerical research conducted to assess the structural

condition of buckled pipes, subjected to both bending and

internal pressure. Fatigue failure under repeated loading and

pipe burst have investigated, through three full-scale buckled

pipe specimens and nonlinear finite element tools as shown in

Figure 1. The maximum strain range from the finite element

computations, and a simple S-N approach gave reasonable

predictions for the number of cycles to failure observed in the

tests. The results of that study demonstrate that under repeated

loading, fatigue failure occurs in the buckled area at the

location of maximum strain range.

Figure 1: Fatigue cracking of buckled pipe under cyclic

loading; test and numerical simulation [10]

The presence of dents has been recognized as a threat to

pipeline integrity, mainly in terms of burst pressure capacity.

Part 12 of API 579-1/ASME FFS-1 [11]code addresses the

assessmentof dents in pressurized components. In general, most

of the available criteria employ a measurement of dent depth to

determine its severity. International pipeline codes usually

consider plain dents non-acceptable if they exceed a depth of

6% of the nominal pipe diameter. In a novel methodology

proposed in Appendix R of the ASME B31.8 code[12], it is

possible to estimate the maximum strain in a dent assuming the

total strain in the dent as the combination of bending strain in

the circumferential direction, the bending strain in the

longitudinal direction and the membrane (stretching) strain in

the longitudinal direction. The value of this strain should be

less than 6% for dent acceptability.

The present study is part of an extensive research program

on the effects of local pipe wall distortions on the structural

integrity of steel pipelines. Preliminary numerical work on the

present research has been presented in [13]. In the present

study, experimental research ispresentedsupported by numerical

simulation in order to investigate the residual structural

integrity of dented steel pipes. A series of six(6) full-scale

experiments on 6-inch pipes of steel grade X52 is carried out.

The steel pipesare initially dented up to different depth values

and, subsequently, they are subjected to further cyclic loading

(bending or pressure) in order to estimate theirresidual strength

and remaining life. Furthermore, finite element analysisare

alsoconductedusing appropriate material models to simulate the

experimental procedure for each type of deformation and

loading case, in order to calculate the stress and strain

distributions at the dented region, so that fatigue life is

calculatedand compared with the experimental results. The

results of the present study can be used forthe assessment of

local wall distortions in steel pipelinestowards efficient pipeline

integrity management.

SPECIMENS AND EXPERIMENTAL SET-UP In this work, six (6) full-scale tests are performed,

consisting of cyclic loading applied on dented pipes. First, the

pipe specimens are dented with a wedge indenter at zero

pressure, and subsequently, they are subjected to cyclic loading

as follows:

four (4) dented specimens are subjected to cyclic

bending until fatigue cracking occurs in the low-cycle

fatigue range,

two (2) dented specimens are subjected to cyclic

pressure loading for a significant number of cycles

(5000 cycles) and then the pressure is raised

monotonically until burst.

The six specimens are 1-meter-longX52 seamless 6-inch

pipes with nominal thickness 4.78 mm

(168.3/4.78),corresponding to D/t= 35. The specimen ends

are capped with thick plates, to enable the application of

internal pressure and the connection with adjacent pipe parts for

the experimental set-up.

In the following paragraphs, the procedure of denting,

cyclic bending and pressure loading are described. It is noted

that denting and cyclic bending has been conducted at the

laboratory facilities of the University of Thessaly, whereas the

two pressure tests have been performed at the facilities of

EBETAM S.A., located in Volos, Greece.

(a) (b)


Figure 2: 6-inch X52 pipe specimens.

(a) (b)

Figure 3: (a) Wedge denting tool and (b)wooden base to

support pipe during bending.

Figure 4: Denting set-up.

Figure 5: Instrumentation of pipe specimen.

Denting procedure Denting is applied on the pipe specimens through the

application of a rounded wedge indenter, shown in Figures

Figure 3a andFigure 4, oriented perpendicular to the pipe axis,

so that a smooth dent is formed on the pipe wall.The total

wedge height is equal to 80 mm, the horizontal length is equal

to 85 mm, and the radius of the rounded wedge is equal to 5

mm. As shown in Figure 3a, the wedge is bolted to the

hydraulic actuator through the use of appropriate steel plates.

During denting, the specimen is supported on a wooden

base (300×280×200) shown inFigure 3b. The top surface of the

base was curved in order to prevent movement of the tubular

specimen. Moreover, to avoid the development of strain and

stress concentration in the pipe at the edge of the wooden base,

the radius of base curve was chosen greater than the value of

the nominal radius of the pipe. Furthermore, the wooden base

was stiffened along both directions (length and height) with the

use of steel rods in order to avoid excessive deformations of the

wood material during the application of denting load.

Prior to denting, uniaxial strain gauges were installed on

several critical points along the outside surface of the specimen,

close to the region where the dent would be developed, in order

to identify local strains during the denting procedure (Figure 5).

Cyclic bending set-up The set-up employed for conducting cyclic bending on the

dented specimens is shown inFigure 6 and Figure 7. The 1-

meter long specimen is connected on either side to two heavy-

walled 7-inch-diameter 650-mm-long tube segments

(193.7/10) made of high-strength steel using a bolted

connection (Figure 7). The entire system is 2.615-meters long,

it is hinged at the two ends (Figure 8a) and it is connected to

the 600-kN-force-capacity hydraulic actuator through a cross-

beam and two wooden clamps (Figure 8b). This corresponds to

a four-point bending structural system, where cyclic bending is

applied through the vertical motion of the hydraulic actuator.

The hinges of this 4-point bending set-up also minimizes the

axial load introduced during the monotonic tests, because of the

constraints applied by the wooden gripsand the supports. Using

appropriately located strain gauges the maximum value of this

at maximum bending load axial forcewas calculatedequal toless

than 8% of the cross-sectional axial yield strength and it is

considered negligible.

Pressure testprocedure The two specimens have been pressurized with the use of a

400-bar-capacity water pump (Figure 9). Cyclic pressure has

been applied with maximum and minimum values equal to 150

bar and 15 bar respectively, at a frequency of about 0.1 Hz.

Following cyclic testing, the specimen was pressurized

monotonically until burst.


Figure 6: Configuration ofthe four-point bending experimental

set-up.

Figure 7: Four-point bending experimental set-up.

(a) (b)

Figure 8: (a) bending specimen hinges and (b) wooden clamp.

Figure 9: Pressure application on specimens.

MATERIAL PROPERTIES AND WALL THICKNESS MEASUREMENTS

Material testing Monotonic and low-cycle fatigue tests were performed on

the X52 steel pipe material to determine material

properties.Strip specimens have been extracted from the

seamless 6-inch pipes,in the longitudinal direction and

machined in accordance with the ASTM E606 standard. The

material stress-strain curve was obtained from tensile coupon

tests and is shown inFigure 10, indicating ayield stress ( Y

)equal to 356MPa, very close to the nominal value, and an

ultimate stress ( UTS ) equal to 554.7 MPaat about 18%uniform

elongation.

Figure 10:Material curve of X52 steel grade

In addition to tensile testing, a total of thirty (30) cyclic tests

were performed on strip specimens withloading ratio R equal

to -1 and 0. In those tests, the hysteresis loops at different strain

ranges were determined and the fatigue ( N ) curve for

thepipe X52 steel were developed. The tests have been

performed in the facilities of FEUP in Porto[14] and the

corresponding fatigue curve can be expressed by the following

Coffin-Manson-Basquin equation.

0.1133 0.4807

0.0102 2 0.333 2N N

(1)


Thickness measurements Prior to testing the specimens,thickness measurements were

obtained using an ultrasonic device at specific points around

several cross sections along the specimen’s length (Figure 11).

A mean thickness value was obtained equal to 5.026mm which

is greater than the nominal value (4.78mm).Furthermore, the

measurements did not show a significant variation of thickness

with respect to this mean value. The value of 5.026mm will be

used in all calculations in this paper.

Figure 11: Steel pipe specimens and thickness measurements

via ultrasonic device.

EXPERIMENTAL RESULTS Table 1, Table 2 and Table 3 show an overview of the

experimental activity. Two values for the dent depth d have

been considered, d D 6% and d D 12% (Table 1). The

first value is equal to the dent acceptability limit used in

practical applications, whereas the second value was chosen

equal to twice this limit. During denting, the specimen was

simply supported by the wooden base while its ends were

totally unrestrained.

The first series of tests consists ofcyclic bending tests on

four (4) dented specimens (SP1d, SP3d, SPd5,SP6d) as shown

in Table 2. The second part of the experimental work consists

of pressure loading on two dented specimens (Table 3).

Denting results

The denting load is applied vertically through the hydraulic

actuatorat a very low speed until a maximum value of dent

depth ofabout16 (for SP1d, SP4d and SP6d) and 27 mm (for

SP2d, SP3d and SP5d), and subsequently, the load was

removed. The remaining dent depth after the elastic “spring

back” is measured about 11mm and 21mm corresponding to

d D 6% and d D 12% respectively.In Figure 12the

denting load-stroke curves for specimens SP5d and SP6d are

presented.The increase of axial and hoop local strains is

illustrated in Figure 13. The dent results are summarized

inTable 1.

Table 1. Pipe denting results.

Specimens

residual

depth

d (mm)

normalized

depth

d D (%)

Maximum

force

maxF (kN)

SP1d 10.245 6.08 69.26

SP2d 20.310 12.06 81.35

SP3d 20.198 12.00 83.29

SP4d 10.593 6.29 69.11

SP5d 20.263 12.04 81.75

SP6d 10.658 6.33 67.39

Figure 12: Load-stroke curves during denting procedure, for

specimens SP5d and SP6d in comparison with the FEA results.

Figure 13: Longitudinal and hoop strain distribution during

denting for specime SP5d (d/D= 12%).

Cyclic bending test results After denting, the locally deformed pipe specimens are

subjected to cyclic four-point bending using the set-up of

Figure 6 and Figure 7, and a displacement-controlled pattern.

The results are summarized in Table 2.

The loading sequence followed in dented specimens SP5d

and SP6d is shown inFigure 14, in terms of the corresponding

load-displacement curves. The two specimens were first bent

beyond the limit loadand subsequently, were subjected to cyclic

loading at two stages. In all this procedure the dent was at the

compression side of the specimen and the load ratio was equal

to 0.1. Following a number of loading cycles of the bent

specimen shown inFigure 15, each specimen failed because of

fatigue crack at the dent “ridge”. The number of cycles to


failure is equal to 1,100 and 950 for specimens SP5d and SP6d

respectively. The crack configuration for specimen SP5d is

shown in Figure 16.

Specimens SP1d and SP3d are subjected to a different

cyclic bending pattern, shown in Figure 17; both specimens

arebentbeyond the limit load [dent in compression].

Subsequently reverse bending is applied, so that the dented

region is under tension, whereas the opposite region of the pipe

exhibits compression. This loading situation results in local

buckling at the opposite to the dent due to excessive

compression. Under cyclic loading, fatigue cracks develop at

the buckle “ridge”.The crack is quite similar to the one shown

in Figure 16.

Figure 14: Load vs displacement curves for specimens SP5d

and SP6d; monotonic and cyclic bending load.

Figure 15: Dented specimen SP5dduring cyclic 4-point

bending.

Figure 16: Pipe wall rupture of specimen SP5d due to cyclic

bending loading at the dent “ridge”.

Figure 17:Load vs displacement curves for specimen SP1d;

monotonic and cyclic bending load.

Pressure tests results Two dented specimens, namely SP2d and SP4d are tested

under pressure loading. The results are summarized in Table

3.Both specimens are subjected to 5,000 cycles of pressure

loading with minimum and maximum valueequal to 1.5 MPa to

15 MPa respectively. No failure or damage has been detected

due to cyclic loading. Subsequently, the specimens are

subjected to monotonically increasing pressure until

burstFigure 18a. The pressure level at which rupture of pipe

wall occurs is 32MPa and 31.6MPafor SP2d and SP4d

respectively. It is interesting to notice that this value is quite

close to the theoretical value of burst pressure estimated from

the following simplified formula 2b UTS

t

D , where σUTS is

the ultimate tensile stress. This result shows that the presence of

a smooth dent on the pipe wall has no effect on the burst

capacity of the pipe.

Furthermore, as shown in Figure 18a, in both cases the

application of internal pressure resulted in a reduction of the

dent size and a “smoothening” of the dented area.In addition,

rupture of the pipe wall occurred away from the dent area, as


shown in Figure 18b, which was also observedin a similar test

on a buckled pipe reported by Dama et al.[10].

(a) (b)

Figure 18:(a) Ruptured specimen due to increased internal

pressure and (b) pipe wall rupture location (different than the

dent location).

Table 2. Cyclic bending ofdented specimens

Specimens d D

(%)

u

(mm) N

SP5d 12.04 13.0 950

SP6d 6.33 13.0 1,100

SP1d* 6.08 16.0 1,400

SP3d* 12.00 16.0 1,200

*Dent in tension (reverse bending).

Table 3. Pressure loading ofdented specimens

Specimens d D

(%)

P (MPa)

N burstP

(MPa)

SP2d 12 13.5 5,000 32

SP4d 6 13.5 5,000 31.6

FINITE ELEMENT SIMULATION

Nonlinear finite element tools have been employed to

simulate the development of the denting and the behavior of the

dented pipes under cyclic bending and pressure loading.The

simulations are conducted with finite element program

ABAQUS 10.1., and the finite element models are capable of

describing large displacements and strains, as well as inelastic

effects rigorously. The pipe is simulated with four-node

reduced-integration shell elements (S4R), which have shown to

perform very well in nonlinear analysis problems involving

large inelastic deformations of relatively thick-walled steel

cylinders. A general view of the finite element model is shown

in Figure 19.

In order to accomplish a good description of the dent

profile, and the corresponding local stress and strain

distribution, the element mesh is rather dense at a two-

diameter-long region, about the middle section of the pipe,

where maximum deformation due to dent is located.

The finite element model consists of five parts; the denting

tool, the pipe specimen, the wooden base that supports the pipe

movement during indentation, and the two stiff pipe

segments(see Figure 6)on either side of the specimen. The stiff

pipe segments are simulated with beam elements of appropriate

cross-sectional and material properties, with simply-supported

conditions are the two ends. The thick end plates were modeled

through a “kinematic coupling” constraint at the end nodes.

Loading is applied at two points of those beam elements,

corresponding to the locations of wooden grips.

To describe inelastic material behavior of the pipe

specimens, a J2 (von Mises) flow plasticity model with

isotropic hardening is employed. The plasticity model is

calibrated through appropriate uniaxial tensile tests on steel

coupon specimens, extracted from the X52 6-inch pipes (yield

stress equal to 356 MPa and ultimate stress equal to 554.7

MPa).

Simulation of pipe denting The wedge-type denting tool is oriented in a direction

perpendicular to the pipe axis, shown Figure 19. For reducing

computational time, only the curved part of the wedge tool,

which is in contact with the pipe, is modeled. The radius of the

rounded wedge is equal to 5 mm identical to the one used in the

experiments. The denting tool is designed as a non-deformable

“analytical rigid” part. Contact interaction is imposed between

the wedge surface and the outer surface of pipe wall.

The geometry of the wooden base,shown in Figure 19, is

similar to the one used in real testing (Figure 3b) with

dimensions 300mm×300mm and the upper surface properly

curved in order to support the pipe model.Contact conditions

are imposed between its surface and the pipe outer surface,

which prevent penetration, but allow pipe uplifting.

The shape of the deformed model is shown in Figure 20.

The pipe model rests on the support plate. During denting, the

two pipe ends are non-restrained, resulting in a small uplifting

of the pipe specimen. Following the experimental procedure,

two values for dent depth are considered in this numerical

study, namely 6% and 12% of the pipe diameter. Those are the

values of the residual dent size after removal of the denting tool

and the corresponding elastic rebound. The dent profile and the

von Mises stress distribution around the dent region after elastic

rebounding are shown inFigure 20. Moreover, force versus

denting displacement curves are compared well with the

experimental curvesfor both cases during the denting and after

the wedge indenter removal, as shown in Figure 12.


Figure 19: General view of the finite element model.

Figure 20: Finite element results for the dented specimen;

distribution of von Mises stress at the outer surface of the pipe.

Cyclic bending of dented pipes Following denting, cyclic bending loading of the dented

pipe is simulated, similar to the experimental procedure. Figure

21 and Figure 22 show the load-displacement curves for those

analyses, and compared fairly well with the experimental

curves shown in Figure 14 and Figure 15. During the numerical

simulation, cyclic bending loading of the specimens is

performed at the loading stages corresponding to the

experimental procedure. At each of these stages, a total of 10

cycles is performed numerically, and the corresponding range

of maximum local strainsin the longitudinal direction with

respect to the pipe axis is measured and depicted in Table 4.

The deformed shape of specimen SP5d is shown in Figure

23, whereas in Figure 24 the shape of specimen SP3d

(subjected to reverse loading) is depicted. In the latter shape,

one may observe the dented area at the top of the specimen,

which is under tension, and the buckled area, developed due to

reverse bending at the opposite side of the pipe wall due to

excessive compression. The buckled area is the one that

corresponds to the maximum local strain range and where the

fatigue crack develops.

Based on the values of local strain range, it is possible to

employ the fatigue curve of the pipe material, expressed by

equation (1), and apply Miner’s rule to obtain a damage factor

fD , as follows:

if

i i

nD

N (2)

where iN is the number of cycles corresponding to

i

obtained from the fatigue curve ( N ) and in is the

number of real cycles applied. Strains are measured at the

outside surface. The results of this analysis are depicted in

Table 4 for the four specimens under consideration. The values

of the damage factor are quite close to 1, indicating a good

correlation between test results and numerical analysis.

Figure 21: Load-displacement curve during bending of

specimen SP3d (reverse and cyclic bending) obtained from

finite element analysis.

Figure 22: Load-displacement curve during bending of

specimens SP5d and SP6d (monotonic and cyclic bending)

obtained from finite element analysis.


Table 4. Fatigue analysis under cyclic bending

Specimen

Cycles

applied

iN

max

(%) fD SNCF

SP1d

500 0.306

1.13

1.239

500 0.246 1.209

400 1.914 11.98

SP3d

500 0.300

0.79

1.214

500 0.248 1.219

200 2.180 13.64

SP5d 500 0.790

1.11 4.184

450 1.750 17.21

SP6d 500 0.603

1.15 3.193

600 1.637 16.10

The value max is the maximum local strain rangein the axial

direction at the dented (or buckled for SP1d and SP3d) (critical)

region, and nom is the nominal strain range due to the

applied loading, calculated through elementary mechanics of

materials, considering the initial (intact) geometry of the

tubular member. For the loading conditions under

consideration, the strain concentration factor is computed

through the measurement of the maximum strain variation in

the longitudinal direction of the pipe, shown in Table 4. In the

case of four-point bending, nom is calculated as follows:

2

2nom

F

E D t

ΔΔ (4)

where F is the range of total transverse load applied on the

specimen, and is the distance between the hinge support and

the point of load application.

The values of strain concentration factor are shown in Table

4. The numerical results indicate that in the last stage of the

bending curve, corresponding to severe deformation of the

pipe, the SNCFcan obtain significant values. This is attributed

to the fact that in that area, the pipe wall is quite distorted, and

the cyclic loading is associated with severe folding/unfolding of

the pipe wall. The calculatedvalues of SNCF are consistent with

those reported by Dama et al [10] for locally buckled pipes.

Figure 23: Deformed shape of specimen SP5d under cyclic

bending conditions obtained from finite element analysis.

Figure 24: Deformed shape of specimen SP3d under cyclic

bending conditions obtained from finite element analysis.

Simulation of pressure loading Internal pressure is applied in the numerical model of the

dented specimen. The results show that with increasing

pressure, the dented profile “smoothens” (Figure 25), the dent

size is decreased, and there is a tendency of gradual dent

flattening, an observation also reported in [10]. Furthermore,

for the range of cyclic pressure applied during the tests

(Δp=13.5 MPa, with a maximum value of 15 MPa), the

corresponding strain concentration factor SNCFis small [A18].

This can verify the fact that the dented pipe is capable of

sustaining 5,000 pressure cycles without failure or other

damage, as observed experimentally.

(a)

(b)

Figure 25: Dent geometry: (a) before pressure application for

d D 6%; (b) after pressure application.

The diagrams in Figure 26present the response of the dented

pipe upon monotonically increasing internal pressure.The

increase of internal pressure results indecrease of dent depth

and the dent profile tends to flatten. The maximum internal

pressure obtained from the finite element analysis isequal to

about Pmax= 33 MPa until convergence of solution is not

possible due to excessive plastification of pipe wall. This

pressure load is well compared with the burst pressure

measured in the experiments.


Figure 26:“Smoothening” of the dent profile with the increase

of internal pressure

The coordinates in the vertical axis of Figure 26 refer to the

position of the top outer wall generator for various pressure

levels. The coordinates are normalized by pipe radius R.

CONCLUSIONS Combining experimental testing and numerical simulation,

the mechanical behavior of dented pipes under pressure and

bending loading is examined. The experimental investigation

consists of six (6) 6-inch-diameter X52-steel pipe specimens,

initially dented at dent sizes of 6% and 12%which is twice as

the maximum acceptable dent depth according to ASME B31.8

provisions (d/D=6%). Subsequently, the specimens are tested

under various loading conditions (bending and pressure). It is

found that cyclic bending on four (4) dented specimens, causes

fatigue cracking, located at the ridge of the deformed area, at

about 1,000 loading cycles. In addition, an accurate finite

element simulation of the experimental procedure (both denting

and cyclic loading), together with the use of an appropriate

fatigue curve for the pipe material, is considered as a sufficient

tool for the prediction of pipeline fatigue life. The results

indicated that the dented pipeswith even a dent equal to

12%Dcan sustain a significant number of pressure cycles while

the level of pressure and the location of the rupture are not

affected by their presence. It is concluded that the fatigue life

and the burst capacity of pipes with local dents,even largerthan

the maximum acceptable limit, have been severely

underestimated by the current provisions.

ACKNOWLEDGMENTS This research has been co-financed by the European Union

(European Social Fund – ESF) and Greek national funds

through the Operational Program "Education and Lifelong

Learning" of the National Strategic Reference Framework

(NSRF) - Research Funding Program: Heracleitus II. Investing

in knowledge society through the European Social Fund. The

authors would like to thank EBETAM A.E. for providing the

facilities for pressure testing, as well as Dr. Abilio M. P. de

Jesus from FEUP for providing the material test data.

REFERENCES

[1] Doglione, R., and Firrao, D., 1998, “Structural Collapse

Calculations of Old Pipelines,” Int. J. Fatigue, 20 (2), pp.

161–168.

[2] Netto, T. A., Ferraz, U. S., and Estefen, S. F., 2005, “The

Effect of Corrosion Defects on the Burst Pressure of

Pipelines,” J. Constr. Steel Res., 61 (8), pp. 1185–1204.

[3] Cosham, A., and Hopkins, P., 2004, “The Effect of Dents

in Pipelines—Guidance in the Pipeline Defect Assessment

Manual,” Int. J. Pressure Vessels Piping, 81, pp. 127–139.

[4] Ong, L. S. 1991, “Derivation of Stress Associated with a

Long Axial Dent in a Pressurized Cylinder,” Int. J. Mech.

Sci., 33 (2), pp. 115–123.

[5] Ong, L. S., Soh, A. K., and On, J. L., 1992, “Experimental

and Finite Element Investigation of a Local Dent on a

Pressurized Pipe,” J. Strain Analysis, 27 (3), pp. 177–185.

[6] Fowler, J. R., 1993, “Criteria for Dent Acceptability in

Offshore Pipelines,” Offshore Technology Conference,

OTC 7311, Houston, pp. 481–493.

[7] Buitrago, J., and Hsu, T. M., 1996, “Stress Concentration

Factors for Dented Tubular Members,” Offshore

Mechanics & Artic Engrg Conference, OMAE, Vol. I, pp.

291–296.

[8] Macdonald, K. A., and Cosham, A., 2005, “Best Practice

for the Assessment of Defects in Pipelines—Gouges and

Dents,” Eng. Failure Anal., 12 (5), pp. 720–745.

[9] Das, S., Cheng, J. J. R., Murray, D. W., 2007, “Prediction

of the fracture life of a wrinkled steel pipe subject to low

cycle fatigue load”, Canadian Journal of Civil

Engineering, 34 (9), pp. 1131-1139.

[10] Dama, E., Karamanos, S. A. and Gresnigt, A. M., 2007,

“Failure of Locally Buckled Pipelines.”, ASME Journal of

Pressure Vessel Technology,129, pp. 272-279.

[11] API 579/ASME FFS-1, “Fitness-for-Service,” 2007

[12] ASME B31.8, 2007, Gas transmission and distribution

piping systems. 2007 ed. ASME.

[13] Pournara, A. E. and Karamanos, S. A., 2012, “Structural

integrity of steel hydrocarbon pipelines with local wall

distortions”, Pressure Vessel and Piping Conference,

ASME, PVP2012-78131, Toronto, Canada

[14] Fernandes A.A. et al.,“Ultra Low Cycle Fatigue of steel

under high-strain loading conditions,” 3rd

Annual Progress

Reportof ULCF-RFCS Project, Porto, 2013

structural integrity of steel hydrocarbon...

Documents